S&P 500
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
S&P 500 Energy sector Window W3
7.14Return
2
S&P 500 Energy sector Window W2
20.38Return (%)
2
S&P 500 Healthcare sector Window W6
4.4Return
2
S&P 500 Healthcare sector Window W5
6.12Return (%)
2
S&P 500 Healthcare sector Window W4
6.41Return
2
S&P 500 Healthcare sector Window W2
16.2Return
2
S&P 500 Healthcare sector Window W1
10.78Return (%)
2
S&P 500 Technology sector Window W6
8.04Return (%)
2
S&P 500 Technology sector Window W3
25.41Return (%)
2
S&P 500 Technology sector Window W2
11.15Return
2
S&P 500 Technology sector Window W1
21.61Return (%)
2
S&P 500 d=98 n=1,253 trading days (full set)
-46.62-ℓ (Negative Log-Likelihood)
2
S&P 500 d=75 n=1,253 trading days
-33.334Negative Log-Likelihood
2
S&P 500 d=50 n=1,253 trading days
-20.553Negative Log-Likelihood
2
S&P 500 2017 (held-out)
56.8Win Rate vs Qwen3-VL-235B
2
S&P 500 900 samples (i.i.d. pool)
0.622KS p-value
2
S&P 500 (1992-2015)
57.8Accuracy
2
S&P 500 Next Day
0.003Mean Return (Long)
2
S&P 500 MALA warm-start 2766 days
99Return Unpredictability Metric (99% CI)
1
S&P 500 (2015-2020)
59.2Directional Accuracy
1
S&P 500 (2010-2020)
58.1Accuracy
1
S&P 500 (2007-2017)
57.3Directional Accuracy
1
S&P 500 (2020-2023)
60.14Accuracy
1
S&P 500 Financial Contagion (test)
—Primary metric
0